Preserving Edge Knowledge Sharing Among IoT Services: A Blockchain-Based Approach

Gaolei Li,Mianxiong Dong,Laurence T. Yang,Kaoru Ota,Jun Wu,Jianhua Li

DOI: https://doi.org/10.1109/TETCI.2019.2952587

2020-01-01

IEEE Transactions on Emerging Topics in Computational Intelligence

Abstract:Edge computational intelligence, integrating artificial intelligence (AI) and edge computing into Internet of Things (IoT), will generate many scattered knowledge. To enable auditable and delay-sensitive IoT services, these knowledge will be shared among decentralized intelligent network edges (DINEs), end users, and supervisors frequently. Blockchain has a promising ability to provide a traceable, privacy-preserving and tamper-resistant ledger for sharing edge knowledge. However, due to the complicated environments of network edges, knowledge sharing among DINEs still faces many challenges. Firstly, the resource limitation and mobility of DINEs impede the applicability of existing consensus tricks (e.g., Poof of Work, Proof of Stake, and Paxos) of blockchain. Secondly, the adversaries may eavesdrop the content of edge knowledge or entice the blockchain to forks using some attacking models (like man-in-the-middle attack, denial of services, etc.). In this article, an user-centric blockchain (UCB) framework is proposed for preserving edge knowledge sharing in IoT. Significant superiorities of UCB benefit from the proof of popularity (PoP) consensus mechanism, which is more energy-efficient and fast. Security analysis and experiments based on Raspberry Pi 3 Model B demonstrate its feasibility with low block generating delay and complexity.